"No hands. No feet." (Via Google)

Now, the Mountain View team is back with a video that allows us to see the software that

makes it all happen. Just watch.

Other cars, cyclists, pedestrians and even construction zones all register in colored

wireframes. But it gets even more specific. (Via Google)

"When the cyclist holds up his arm, our software detects the hand signal and predicts his movement

into our lane."

The company's vehicles have now logged a reported 700,000 autonomous miles. And lately, Google

says, it's focused its attention on driving in Mountain View rather than on simple-to-navigate

freeways.

Chris Urmson, director for the self-driving car project, writes the variety of obstacles

on a typical city street can leave any driver frustrated. "A self-driving vehicle can pay

attention to all of these things in a way that a human physically can't—and it never

gets tired or distracted." ​(Via Google)

Take another look. The software is registering all of those obstacles simultaneously — likely

more than you could say for your average human driver. (Via Google)

​The Atlantic, which got a ride-along, points out people still serve a purpose in Google's

future machines. During the experience, the driver reportedly needed to take manual control

of the car twice.​ But "it's clear, now, that Google's driverless cars are a real thing."

That said, Mountain View isn't Everytown, USA.

"A writer for GigaOM says, "I would also add that while navigating Mountain View roads

may be tricky, talk to me after Google cars have mastered the potholes, rotaries and creative

drivers of Boston."

That's a feat Urmson says the cars are still working toward. He says the vehicles still

need to master Mountain View before they can "tackle another town." (Via Google)

CNET takes a step back from the tech, saying Google is spearheading the self-driving movement,

but "It's not yet clear how exactly the company hopes to profit from it."

Maybe we'll know more on that two years from now.